Janet M' Wang - PowerPoint PPT Presentation

1 / 18
About This Presentation
Title:

Janet M' Wang

Description:

Janet M' Wang – PowerPoint PPT presentation

Number of Views:16
Avg rating:3.0/5.0
Slides: 19
Provided by: carl290
Category:
Tags: algo | janet | wang

less

Transcript and Presenter's Notes

Title: Janet M' Wang


1
Extended Krylov Subspace Method for Reduced Order
Analysis of Linear Circuits with Multiple Sources
  • Janet M. Wang
  • Tuyen V. Nguyen

University of California at Berkeley,
CA94706 IBM Austin Research Laboratory ,
Austin, TX 78758
2
Outline
  • Introduction
  • Power grid analysis
  • Multiport reduced order modeling of the power
    grid
  • Input dependent explicit moment matching
  • Extended Krylov subspace method
  • Incremetal orthogonalization algorithm
  • mulpoint extension
  • Practical considerations
  • Scaling
  • Error control
  • Experimental results
  • Conclusions

3
Introduction
4
Introduction
5
Outline
  • Introduction
  • Power grid analysis
  • Multiport reduced order modeling of the power
    grid
  • Input dependent explicit moment matching
  • Extended Krylov subspace method
  • Incremetal orthogonalization algorithm
  • mulpoint extension
  • Practical considerations
  • Scaling
  • Error control
  • Experimental results
  • Conclusions

6
Power grid analysis
  • Multiport reduced order modeling of the power
    grid
  • Input dependent explicit moment matching

7
Multiport reduced order modeling of the power
grid
Krylov subspace
The longer the u vector, the bigger the reduced
order matrices
8
Input dependent explicit moment matching
9
Input dependent explicit moment matching
10
Input dependent explicit moment matching
11
Input dependent explicit moment matching
12
Outline
  • Introduction
  • Power grid analysis
  • Multiport reduced order modeling of the power
    grid
  • Input dependent explicit moment matching
  • Extended Krylov subspace method
  • Incremetal orthogonalization algorithm
  • mulpoint extension
  • Practical considerations
  • Scaling
  • Error control
  • Experimental results
  • Conclusions

13
Extended Krylov Subspace method
Input source moments
Congruence transform matrix
14
Incremental Orthogonalization algorithm
At
15
Incremental Orthogonalization algorithm
At
16
Incremental Orthogonalization algorithm
At
For j1k-1
17
Incremental Orthogonalization algorithm
  • Theorem 1 The orthonormal vectors generated
    by the above algorithm span the same space as the
    original moment vectors

18
Multipoint Extension
In order to capture higher frequency content of
the input waveforms
Write a Comment
User Comments (0)
About PowerShow.com